Our planet monitored in millions of megapixels

Published: 11 January 2018

It is vital to monitor changes in surface water worldwide to understand how our planet is changing as a result of human activities or climate change. Gennadii Donchyts (a Deltares expert in remote sensing) has investigated existing ways of detecting surface water, and developed new algorithms, in his thesis 'Planetary-scale surface water detection from space', which he defended today at Delft University of Technology.

Water sources under pressure

The monitoring of surface water is becoming increasingly important because water sources are under pressure from economic sectors such as industry, agriculture, energy, tourism and domestic use. Furthermore, water availability is decreasing, driving more regions into water insecurity. Increasing volumes of satellite data allow the monitoring of surface water at higher resolutions. At the same time, more efficient and accurate methods are needed to monitor surface water on the global scale. In his thesis, Gennadii studies automated ways of detecting surface water using radar satellite images.  He explores a number of existing methods and introduces new algorithms for highly accurate estimates of the area of surface water and any changes. These methods are applicable on the local to global scales.

Deltares Aqua Monitor

To map out changes in surface water worldwide over the past thirty years, Genna developed the Deltares Aqua Monitor using a multi-petabyte archive of millions of Landsat satellite images collected by NASA and USGS in recent decades. Combining datasets from all the freely-available satellite sensors required the harmonisation of a large number of satellite images, but also enormous computing resources.  Most of the analysis presented in Genna’s study therefore involved using the Google Earth Engine platform.

Reservoirs

Another area addressed by Genna’s thesis is information about the number and size of the world’s reservoirs, and changes in them. Reservoirs, which are often used to generate electricity or flood risk management, play a major role in river water management. There are more than a million reservoirs in the world and the number continues to rise.  Information about water storage and changes in these reservoirs is sometimes lacking; sometimes it is kept back for political or economic reasons. The reservoirs play a major role in limiting the effects of droughts and floods, and information about them is important for accurate hydrological forecasts. So observations from space are essential to monitor changes. In his thesis, Genna demonstrates that reservoirs can be accurately mapped using satellite images, even when there are clouds or ice. He developed a new approach using older satellite images to learn from the past and fill in the missing pixels. This new algorithm gives us a better picture of how reservoirs are being managed and allow stakeholders to make better use of them.  Gennadii: ‘The ability to manage such large amounts of satellite data in the cloud and to zoom in on the details will give a new impetus to the earth sciences, something we didn’t think was possible ten years ago’.

New users

The availability of analyses made on the basis of satellite data will result in new users, an area that Genna also covers in his thesis. On the very local scale, the general public can now make better decisions without expert help about the impact of coastal erosion on their homes. On the regional scale, downstream countries can use annual measurement data to see whether their upstream neighbours are extracting more water. Finally, on the global scale, international organisations such as the United Nations International Strategy for Disaster Reduction can monitor whether enough storage capacity is being created to reduce flood risks.